Bid optimisation methods for real-time bidding in online display advertising.

---

CONTRIBUTERS:

• Alexander Cowen-Rivers (GitHub)
• Lynray Barends (GitHub)
• Tim Warr (GitHub)

---

Instructions

Pre-processed Dataset

1. Download the files https://drive.google.com/drive/folders/165CDcG3pTd07-XUFon5M2cWS47hvLMnR?usp=sharing
2. You must then create a subfolder Data/.
3. Split the downloaded data into the subfiles Data/train.csv, Data/validation.csv and Data/test.csv.

Full Dataset

1. To get the full dataset, download the data dump (6.6 GB zip) from http://data.computational-advertising.org/.
2. You must then create a subfolder Data/.
3. In a seperate notebook you will then need to split the downloaded data into a pandas dataframe, with columns for the user profiles, click information and payprices.
4. Split the data into the subfiles Data/train.csv, Data/validation.csv and Data/test.csv.

For:

• Reports see GROUP / ACR / LB / TW
• Data Exploration see ACR / LB / TW
• Basic Bidding Strategies see this
• Linear Bidding Strategy see this
• Indiv Bidding Strategies see ACR / LB / TW
• Group Bidding Strategies see - - - - NEURAL / MULTI-AGENT (By ACR, located after the single agent experiment)- -

---

Dependencies

• tbc